7 EMIT Sloan School of Management

Transcription

1 BALANCING FLEXIBILITY AND FORMALITY IN PRODUCT DEVELOPMENT AT A HIGH GROWTH TECHNOLOGY COMPANY by Gary S. Koerper B.S. Electrical Engineering Purdue University, 1994 Submitted to the Sloan School of Management and the Department of Electrical Engineering and Computer Science in partial fulfillment of the requirements for the degrees of Masters of Science in Management and Master of Science in Electrical Engineering and Computer Science In conjunction with the Leaders For Manufacturing program at the Massachusetts Institute of Technology May 1998 MassachusettsI stitute of Technology, All Rights Reserved. Signature of Author, 7 EMIT Sloan School of Management Departmen ielectrical Engineering and Computer Science Certified by Certified by Professor Michael Cusumano MIT Sloan School of Management Thesis Supervisor Certified by Accepted by Professor Alvin Drake Department of Electrical Enineerin, and Computer Science *is Supervisor Arthur C. Smith, Chairman, Committee on Graduate Students Department of ElectriC Engineering and Computer Science Accepted by Lawrence S. Abeln, Director MAS Abi,,uti.tS TITUTE Sloan Master's Program OF TECHNOLOGY JUN LIBRARIES

2

3 BALANCING FLEXIBILITY AND FORMALITY IN PRODUCT DEVELOPMENT AT A HIGH GROWTH TECHNOLOGY COMPANY by Gary S. Koerper Submitted to the Sloan School of Management and the Department of Electrical Engineering, in partial fulfillment of the requirements for the degrees of Master of Science in Electrical Engineering and Computer Science and Master of Science in Management. Abstract In the competitive market of wireless telecommunications, time to market for new products can be a key strategic advantage. A short cycle-time product introduction process allows companies to respond more quickly to customer's needs, incorporate a greater span of new technologies and improve product functionality and cost. This thesis presents a study on how balancing flexibility and formality in a New Product Introduction (NPI) process can increase time-to-market for large scale technology products. This thesis begins with a background on the technology of wireless communication and Qualcomm Inc, the wireless equipment provider this thesis is developed in partnership with. Chapter 3 discusses the results of a post mortem conducted on the Qualcomm's existing New Product Introduction (NPI) methodology. The results of the post-mortem are broken up into three categories: general working environment, communication between functional groups, and process concerns. Chapter 4 uses the results from the post-mortem to identify linkages between the design and manufacturing groups that are critical to balancing flexibility and formality in the NPI process. An extensive analysis is conducted of this critical link, including a discussion of Microsoft's "synch-and-stabilize" process used to address this issue in software development. Finally, Chapter 6 presents specific recommendations to optimize the link between design and manufacturing for balancing flexibility and formality in the NPI process. The thesis concludes with a discussion covering issues relevant to the implementation of the recommendations made in Chapter 6.

4

5 Acknowledgments I would like to take this opportunity to express appreciation and thanks to individuals and organizations who have helped me pursue this thesis, my education at MIT and my career over the last two decades. MIT's Leaders For Manufacturing program and the Qualcomm corporation have been instrumental in the development of this thesis. The LFM program has afforded me the opportunity to expand and challenge my beliefs and philosophies about leadership, technology and the future of American manufacturing. Chris Ross and his staff at Qualcomm invited me into every part of their organization, and were critical to my success and learning during the time with them. I would like to thank my Electrical Engineering/Computer Science advisor Al Drake for his wisdom and guidance through the 2 year maze of the LFM program. I also want to thank my Management advisor, Michael Cusumano for providing me with an intimate knowledge of technology companies during my internship. This work was partially supported by funds made available by the National Science Foundation Award #EEC (Technology Reinvestment Project #1520). I would also like to thank my fiance, Tiffany Marie Anastasia Archer for her unconditional love and support at every critical point in my adult life and career.

6

7 Table of Contents 1.0 Introduction... 9 The Strategic Im portance of New Product Introductions Goals of the Thesis Project Background Com pany Technology Anatom y of a phone call Existing New Product Introduction Methodology Phases Phase 1: Product Definition Phase 2: Product Design Phase 3: Product Prototype Phase 4: Pre-Production Phase 5: Production Release Control of Docum entation Post M ortem on Previous Product Introductions Process of Inform ation Gathering General W orking Environm ent Start-up Culture Power Hierarchy Com m unication between functional groups Integration of Operations and Design Integration of M arketing Process Issues Lack of Form ality Project Scheduling Reliance on Operations General Conclusions The Balance Between Flexibility and Formality NPI process - Critical Link

8 4.2 The Present Organizational Structure The Engineering Planner The NPI Manufacturing Engineer The Manufacturing Development Lab Definition of Optimization of the NPI critical Link The Sync and Stabilize M odel Sync and Stabilize and the Microsoft Corporation Sync and Stabilize at the Qualcomm Corporation Recommendations for NPI Improvement High Level Organization - Phase Review Management Manufacturing Requirements Document Supply Chain M anagem ent Functional Design and Development Im plem entation C onclusions B ibliography

9 1.0 Introduction The Strategic Importance of New Product Introductions In the past decade, technology organizations have begun to realize the importance of time-tomarket. Technology is evolving at such a pace that design, development and manufacturing are finding it difficult to keep up with the demands placed on them by growing customer expectations and increasing competition. Continuously shrinking product life cycles are making time-to-market a critical success factor for programs and for companies as a whole. As technology products have market lives which are often less than a year, and strong competition is creating extremely thin margins, being late to market can significantly reduce the success and lifetime of a product. Goals of the Thesis Project This thesis represents the culmination of six months work at a wireless telecommunication equipment provider associated with MIT's Leaders For Manufacturing (LFM) program. The project was originally conceived of as an opportunity to improve the New Product Introduction process within one of the company's divisions. As this was a joint effort, the thesis had to satisfy both the academic requirements of MIT as well as the company's need for a NPI process improvements, implemented into their systems. With this in mind, the thesis project was completed in three major phases. The first was performing a broad evaluation and analysis of the established NPI process. The second was to continue with more detailed research into a fundamental linkage within the greater process that was critical to the success of the overall NPI effort. The third was to implement specific well-defined improvements to the NPI process. 2.0 Background 2.1 Company Qualcomm is a relatively "young" rapidly growing wireless telecommunications company in a hyper-competitive industry with many large and well-established competitors. Though the organization has been in existence for over a decade, only recently has it entered the wireless

10 communications market. Within the wireless technology market, this company is considered to be a relative newcomer, having only produced and sold products for approximately three years. The company utilizes a standard matrix based management system, with operations, engineering, program management, product management and marketing in separate functional groups. Product based teams are coordinated by dotted-line assignments to cross-functional program managers. Though the structure is not unique, the corporate culture is. The environment is that of a "startup" firm, implying incredible growth in both revenue and employee headcount, with relatively few entrenched business practices and procedures. The general philosophy has been to employ very intelligent people and place them into an environment that fosters creativity and drives products to the market quickly. Based on the revenue growth over the last three years, this formula has been very successful. However, this tremendous growth has added to the complexity of the corporation which demands more formal business processes. The employees recognize the necessity for this change, but adding structure and formality to the creative engineering core is a difficult change, especially when it is believed that their entrepreneurial spirit is the basis for success. The company is engineering driven, in that both technical and business decisions are informally made and directed by engineering or by the Chief Executive Officer, who has always been a influential technical force in the corporation. As engineering often takes the lead in determining next generation products, marketing and other functional groups are forced to be responsive to those decisions. There is some movement within the organization to make the company more marketing driven, although that change has been slow to take hold. In general, these characteristics are systemic of a "start-up" corporation attempting to transition to the next stage of its evolutionary life. 2.2 Technology There are several divisions of this corporation that produce a variety of wireless communication products. This thesis was written in conjunction with the Wireless Infrastructure Products Division (WIPD), which designs, develops and manufactures the support systems for wireless communication [See figure 1]. A brief explanation of this technology would be prudent.

11 2.2.1 Anatomy of a phone call This is a rudimentary explanation of a very complex system that BSC supports wireless communication. There are three different communication links that are established when a wireless phone Of call is placed. 1)The outgoing call placed on a cellular phone TS (subscriber unit) will first communicate to a Base station Transceiver Sub-system (BTS) via an wireless radio link. This BTS is Figure 1 responsible for communicating with multiple cellular phones simultaneously and is the primary component of the wireless infrastructure system that determines the overall capacity of the wireless network. Though there are many factors that determine the individual call capacity, a single BTS can usually handle upwards of 65 simultaneous phone calls. A network of BTS cells are deployed to cover an entire geographic region and demographic market to service all subscribers in that location. Figure 2 shows how multiple BTS's (designated by the white circles) are positioned to provide coverage for the entire North County region of San Diego. These BTS's are strategically placed to cover high caller density areas (city centers) and routes of transportation (highways) with secondary systems covering lower-priority areas. 2) The BTS takes Figure 2 the conversation encoded in the radio signal from the subscriber unit and passes it along to the Base Station Controller via a digital backhaul. There exists multiple BTS's that simultaneously transmit signals back to a single BSC. 3) The BSC passes the signal along the Public Switched Telephone Network (PSTS) via a digital Trunk line. The BSC also provides electrical and signal

12 isolation as well as switching, billing and feature capabilities. A signal traveling to the cellular phone is communicated in reverse fashion. In the creation of a cellular system for a geographic location, multiple BSC's and the corresponding BTS's are located within that region to create the wireless infrastructure to support cellular communications. The art of determining where to place the BTS's and BSC's in order to optimize capacity and coverage is critical to planning the wireless network for any given city. The combination of the BSC's and BTS's represent the wireless infrastructure products. US Wireless Infrastructure Market Qualcomm Inc. is new player in the wireless infrastructure products market, having entered Lucent 36.2% 37.2% Ericsson 28.0% 23.2% competition as of As of 1996, Qualcomm Motorola 19.8% 14.5% comprised only 1.5% of the $5.3B dollar domestic Nortel 15.3% 19.2% Nokia 0.0% 3.7% wireless infrastructure business [See Figure 3]. Being Qualcomm 0.0% 1.5% Others 0.8% 0.6% a new entrant in this market, their product portfolio is Figure 3 relatively small compared to the other big players such as Motorola, Nortel and Lucent. Presently, Qualcomm has two different BSC systems, called the QCore products, and six different BTS's, called the QCell products. Products are differentiated by capacity (number of simultaneous calls that can be handled by the unit), operating frequency(cellular or PCS band), installation location (inside or outside), and specific customer requirements. New products are launched according to the dramatically changing market and technology. Generally speaking, a completely new design for either a BSC or BTS, including changes to the silicon chips internal to the system would take on the order of 2-3 years. A minor revision of a product for reintroduction into a different market niche may take on the order of 10 to 15 months. Within the next 2 years, approximately 5 new products are expected to be launched, which will either add to the product portfolio or replace existing products. The purpose of this thesis is to propose ways of decreasing the cycle time for product introduction as a means to increase the competitiveness of Qualcomm. 2.3 Existing New Product Introduction Methodology Though the absence of an official New Product Introduction (NPI) methodology is widely accepted within the corporation, an unofficial system has evolved. The organization generally hires only experienced people, many having developed products at competitive telecommunication companies. As such, an NPI process has naturally developed based on tribal

13 knowledge and past experiences from other organizations. As the process is informal, employees with different backgrounds evaluate their positions, roles and responsibilities inconsistently from each other. Moreover, the boundaries of those duties in relation to other employees and the rest of the organization are unclear, and therefore prone to either redundant action. The system has been successful thus far because generally speaking, the individuals are very pro-active and feel a strong connection between their job and the overall success of the company. However, repeated long hours over the last several years and the rapid growth of the company have left the workforce looking for more formality and structure to the NPI process Phases Figure 4 below shows that there are five phases to the NPI process at Qualcomm. This is not to imply that this unofficial process is formalized to the point where there are distinct bounds on any of the phases, or that all employees are keenly aware of the existence of the individual. Figure 4 only attempts to describe the general approach to NPI that the company and its functional groups have taken. Phasel: Definition P h a se 2 :D e si gn Time Phase 3: Mock-up/Prototype Process Flow Phase 4: Pre-Production Phase 5: Production Figure 4 Phase 1: Product Definition As the organization transitions from being engineering-driven to more marketing-driven, the process by which products are defined is in flux. Previously, VPs of Technology or in some cases the CEO would dictate what products would be developed based on their knowledge of the market, trends and technology. As the company changes to become more market oriented, the product definition has begun to take on a more standard process as shown in Figure 5. This flow chart of information attempts to demonstrate the more logical procedure that the organization takes to determine what products should be included in the portfolio of the company. However,

14 this is not to imply that the company is this structured, but this is merely a general map of a very loose action and decision process. 1) From market research and information from existing and potential customers, marketing determines a potential product for the Wireless Infrastructure Group. Typically, this idea for a new product is in reference to bids that marketing/sales have made to potential customers for new business. 2) Product management takes the high level product concept and Sales/Marketing Identifies a Potential for New Product Product Management roughs out Product Definition based on dialog with Sales/Marketing sales 2-3 develops year Product Management Develops a Salecs forecast dev flops 2-3 year Product Requirements Document with roduct input from systems engineering and begins to derive a functional map of the..o Program Management product itself. 3) Armed with the functional definition of the product, sales develops a revenue forecast, as product management develops a more technical specification with the help of program management and systems engineering. From this technical Product Requirements Document (PRD), Product Management Develops Pro-Forma P&L around Business Case Product Management sells Executive Management on product idea Program Management create manpower, resources, cost and schedule estimates Go/No Go Decision development costs and schedules are estimated. 4) Product management then Figure 5 develops a pro-forma P&L with which they sell the product to executive management. Once the product concept has been accepted, program management and engineering begin phase 2, Product Development. This very informal process is completely circumvented by those products that are defined and driven by the Engineering group or by the CEO. Unfortunately, though phase 2 may have begun, the Product Definition phase has yet to end, as shown in Figure 3. It has been mentioned that the original idea for the product came from a sales bid for business with a potential customer. As the bidding process with customers continues, the definition of the product changes. For example, only when a contract is signed with the prospective customer is a network plan established for that individual market, dictating the size

15 and specific functional requirements of the system. As such, many features of the system remain in flux, even up until shipment. This process appears to be representative of the entire wireless infrastructure market, and is not necessarily specific to Qualcomm. Phase 2: Product Design The design of either a BTS or a BSC is a very complex process including electrical, mechanical and software engineering. In association with the actual product itself, is the corresponding testing systems which have a similar level of complexity and integration. Practicing concurrent engineering, the company designs and develops the electrical Circuit Card Assemblies (CCA) simultaneously with the mechanical racks and backplanes that will be used to connect and support them. Development on the testing systems are also done in parallel to the development of the actual product. It is the responsibility of program management to coordinate the design and development efforts in engineering with manufacturing, training and field engineering. The longest lead-time for individual components for these systems has historically been cables. On any given system, there are over 300 cables of varying types, connectors and lengths which connect the individual subsystems and the various racks of electronics. In an attempt to define the cables as early as possible, manufacturing is extremely proactive in creating a mock-up system in parallel to the product design effort of the design engineers. This non-functioning system is used to determine the routing and lengths of any cables defined by the design engineers. NPI manufacturing engineers will usually begin development of the Mock-Up as soon as the base rack structure is designed, and the location and design of a few subsystems is determined. This is not meant to be a fully functional system, and in many cases, cardboard cutouts for shelves, components, etc. are used to determine cable specs. This Mock-Up is usually transitioned into the first functioning Prototype in phase 3. However, the point of transition from phase 2 to phase 3 is not defined by product maturity, and in almost all cases Prototyping begins before Product Design is completed. In fact, the design of the product continues to evolve and change well into the phase 4, Pre-Production. Phase 3: Product Prototype In this phase, manufacturing builds a few (anywhere from 1 to 10) functional units for internal hardware and software development. This is done primarily in the Manufacturing Development Lab (MDL), removed from the manufacturing floor, and away from the constraints of formal

16 manufacturing processes. The purpose of this phase is to prove out the design, and work through problems and issues while the design and corresponding processes are still flexible. Phase 4: Pre-Production The purpose of the Pre-Production phase is to prove out the manufacturing processes. At the beginning of phase 4, preliminary work instructions (documentation dictating how the product is to be put together) and shop floor control routes (map of the product's movement across the production floor) are supposed to have been created and instituted by the beginning of the phase. All manufacturing processes are supposed to be finalized and validated before entering phase 5, Production. At least one of the Pre-Production units will be shipped to an external customer as a Beta trial unit. Phase 5: Production This phase should represent the point in the product's life where both the design and manufacturing processes are mature and proven. This includes all product integrity testing intended to prove out functionality as well as regulatory approval. In reality, the product and processes continue to evolve and change through to Production phase Release Control of Documentation In the development process, the revision levels of a document attempt to communicate and control the maturity of the drawings of parts, components and higher level assemblies of the products. There are two levels in the revision system. 'X' and '-.' Under the rules of the release system, a document is considered to be at Revision X, if the cognizant engineer has authorized that document to be under the control of Configuration Management (CM). This entails the engineer signing off on the document and turning it over to CM for archiving purposes. The advantage of having documents (and therefore parts) at Rev X, is that changes can be made quickly by the cognizant engineers, as only their signature is needed to enact changes to the documents under control by CM. The second level of the Revision System, Rev -, represents a much greater control over documentation. In order for documents to reach the Rev - level, they must be voted on by the Change Control Board (CCB), which includes the cognizant engineer, Project Manager, Program Manager, Manufacturing, Quality as well as other representatives whose inclusion in the CCB are at the discretion of the CM group. The intention of the CCB and the Rev - release level on documents is to control the risk associated with changing the

17 documents (and hence the product) arbitrarily. Notification of changes are made via and informally through "hallway conversations," which is considered a significant and critical communication medium within the organization. According to corporate ISO policies, an entire product is supposed to be released to Rev - before it enters full production with unlimited release to external customers. Corporate practice is much different, as several products studied in the post mortem were launched into production while much of the product remained at Rev X. 3.0 Post Mortem on Previous Product Introductions In an attempt to better understand the organization and previous successes and failures for this thesis, post mortems were conducted on three different products launched over a period of thee years. 3.1 Process of Information Gathering Thirty employees with experience in previous product launches were interviewed. Interviewees included members from program management, product management, production, materials, quality, executive management and marketing. After collecting the information, the opinions and thoughts were scrubbed by using TQM practices including a KJ. 1 In this process, the individual thoughts and ideas were categorized into several general groups. These groups included general working environment, communication between functional groups and process concerns. 3.2 General Working Environment Start-up Culture Almost unanimously, the employees described Qualcomm as an engineering driven organization with a very unique culture. They directly attribute the company's success to the start-up culture that their founder, Irwin Jacobs, has instilled in the corporation and fights to keep alive. Despite the tremendous influences to grow the organization, human resources has apparently resisted the need to fill positions with whomever they can find. Instead, they have historically hired only the best and the brightest of the engineers on the market, and gives them an equity stake in the company. This has served to create a very strong engineering-entrepreneurial culture with a lot of freedom and creativity afforded to the engineering community. This sort of organization has been very successful in the past, but the employees recognize that the company is at a crucial stage in its evolution, where they must become more mature with established business processes and formal methodologies.

18 3.2.2 Power Hierarchy There was a general feeling among the parties interviewed that a power hierarchy existed among the functional groups. Design engineering was perceived to be in control of projects, with manufacturing, program management and marketing playing subordinated roles. One of the program managers responsible for a product launch summarized it as follows. My litmus test for doing anything on a product or project is whether I can get the support of design engineering. If they buy into the idea, it has a good chance of success. If design engineering does not buy into the idea, there is pretty much no chance of it getting off of the ground. Not only did design engineering tend to wield most of the control, there was also the concern among the groups that design tended to drop support of the product much too early in the product launch. Historically, by the time the latest product was just entering the pre-production and production phases of the NPI effort, a newer and more exciting product was being defined by marketing. Design resources tended to be pulled off of the existing product too early, for the sake of staffing up for the new product coming down the pipe-line. As a result, other organizations felt that design was not very supportive of their needs in the latter stages of product launches. For the most part, operations was highly regarded across the entire organization and throughout all levels. The general impression among other groups is that operations people are motivated, proactive individuals who take ownership of their problems and do what it takes to pull off miracles in getting the product to market. Almost all of the people within operations have extensive experience in manufacturing from other organizations including DEC, Motorola, General Dynamics, Unisys and others. Having come from a variety of backgrounds, these individuals all had different approaches to their jobs and philosophies on the roles and responsibilities of their respective assignments. 3.3 Communication between functional groups Integration of Operations and Design There was a tremendous effort on the part of operations to be part of the design process. There were no formal lines of communication across functional groups, so operations became very aggressive in how they "embedded" themselves into the design process. In order to aid in the

19 communication between the design engineers and operations, the manufacturing engineers and the materials support organization were staffed in the same building as the design groups. This served to close the communication gap between the designers and the individuals in operations that were directly in contact with them. Operations did well is to integrate themselves informally upstream into program management and the design activities so as to push engineering for the deliverables that manufacturing needed to get the product launched. None of this was by formal process, but almost entirely because operations was very proactive in getting what they needed to accomplish their goals. Operation's attempt to move upstream into the process of design engineering was continuously met with mixed reactions from the engineering community. There were apparently circles of influence in the design groups which dictated how the design engineers reacted and behaved to manufacturing's role in phase 2: Product Design. If the VP of Technology or the Engineering Director in charge of the design was sympathetic to the need for operation's involvement, integration upstream became much easier. On the products where the leaders of the design staff were less understanding, operations was forced to find more creative ways to get their job done. One of the operations program managers in charge of a New Product Introduction spoke to this issue: The design community at Qualcomm can tend to have a very myopic view of product development. They believe that all that matters is how long it takes them to get the design done and working. They often don't see the need for manufacturing's involvement in a project, and feel that we just slow them down. We may force them to go more slowly in terms of just the design phase, but our involvement decreases the total time-to-market of the entire project. We are getting better at this Integration of Marketing There exists a large communications gap between marketing and both design engineering and operations. As was mentioned previously, the perceived organizational hierarchy places marketing on the bottom of the food chain at Qualcomm. Marketing's relative position in the organization leads to disconnects and mistrust with the rest of the corporation. On one product launch, marketing had committed to sell a specific feature which it felt to be a non-issue and easily developed. In reality, that specification represented a complete redesign of major systems and months of development time on the part of both design and manufacturing. Marketing's

20 lack of connection with the technical competence of the organization led to a decision that postponed the ultimate delivery date of the product by over a month. Their lack of power in the structure seemed to be amplified by the difficult position that they play in attempting to translate a very tumultuous market to product definitions and forecasts that drive design and operations. Marketing is responsible for developing forecasts for new products yet to be launched, as well as established products being shipped into the market. Because of the volatility of the bidding process necessary to win contracts, and the underlying volatility of the market, it was very difficult for marketing to develop accurate forecasts, making operations and design engineering even more skeptical of the information from marketing. This reinforced the relational gap between the functional groups, and caused the communication to become more dysfunctional. This marketing disconnect is not necessarily unique to Qualcomm, and is often found in a wide variety of high-tech companies. 3.4 Process Issues Lack of Formality There are no formal processes in place to direct or coordinate design, quality, manufacturing and procurement. When asked the question of what the New Product Introduction methodology was at Qualcomm, the common answer was "We have an NPI process?" Qualcomm is still developing its own methodologies, despite the fact they have over $2B in sales with over 9000 employees. Consequently, there are no formal lines of communication, and the exact roles and responsibilities of respective employees are not clearly known. This seems to work fairly well because the people in the organization are very proactive and work to cover all issues on a product launch, but the lack of structure and formality results in everyone working twice as hard as they should. Design engineering was the main opponent to adding any formality or structure to the product development process. They felt that such constraints would inevitably turn into a "Gated" development philosophy, slowing down their individual development effort. Several of the interviewees from manufacturing jokingly referred to the corporate philosophy on a gated product development process. If you mention the word "Gate" in this organization, you will be thrown out of Qualcomm.

21 This had an adverse effect on the rest of the organization where jobs had to be aligned with the design group's efforts and the evolving product. As there were never any points where the design would be frozen, or any gates in the process that would allow the other parts of the organization to sync-up efforts with design, many of the other groups tended to lag behind the development effort. The rest of the organization would have to scramble to stay abreast of the constantly changing design. It was described before that Qualcomm used such terms as Prototype, Pre-Production and Production to theoretically describe the state of a product or project. Unfortunately, there was not common agreement as to the real meaning of these terms across functional groups, or even internally within the vertical organizations. As design would dictate changes to the product even up to high volume manufacturing, Production or Pre-Production did not have much significance as far as the maturity of the product or processes. This ambiguity caused many problems within the manufacturing organization Project Scheduling In Qualcomm's process for defining products, detailed in Figure 4, there existed a breakdown in the flow of information that negatively affected operations. In the beginning of each product launch, program management was responsible for identifying resource needs from engineering and manufacturing in order to create a cost model and delivery schedule for the project. All told, the headcount for a larger BSC effort may include around 50 design and manufacturing engineers over a two year period. This is further broken up into smaller teams of 3-5 persons each, for electrical design, mechanical design, tester development, software and manufacturing. This represents only the those people directly staffed to a single product, as total headcount dealing indirectly with a product would be much more. Program management would work with engineering to understand design needs and conflicts, but would often leave operations out of the total schedule creation process. Program management would take their best guess as to the needs for manufacturing resources and time. Only until this budget and schedule had been accepted by executive management would operations get the opportunity to see and react to the schedule. By that point, they would be powerless to push back on a delivery schedule that had already developed momentum and expectations in the upper levels of the company and within marketing. In one specific case, operations was particularly frustrated with the schedule that was created. An operations program manager spoke to this.

22 Even from the beginning, we knew that the schedule was completely unrealistic, but nobody believed us. We fought with program management and marketing to convince them that the product would not be done, but they would not slip the schedule. Everyone expected us to pull off miracles, because we always did it before. The expectations for us to do the impossible became so great that ultimately the team dynamics started to break down within manufacturing, because we knew we couldn't do it but no one would listen. Ultimately, we were not able to meet the really schedule that was set for us, but we still felt successful because we met the schedule we had created at the start. Marketing and program management, however, felt that manufacturing was unsuccessful, having not met the original schedule that program management had created. This lack of alignment of goals and expectations between the groups occurred on other development projects as well Reliance on Operations Operations had little control on project schedules, which were ultimately determined by marketing and program management. Being at the end of the product launch process meant that operations was often held accountable for making up time in the schedule that the design group had lost. On several occasions, despite the fact that design engineering slipped their schedule back due to unforeseen problems, manufacturing was still expected to launch the product on the pre-determined date. This served to force manufacturing to work extremely hard to "pull off miracles" and get the product to market. However, this served as a reinforcing dynamic loop for the design community. After the second product launch, operations was expected to always be able to execute miracles which became very detrimental to the attitude of the operations group. An operations program manager spoke to this: Operations is clearly under the gun. Design will consistently slip their schedule for delivering information and designs to us, but we are still expected by the organization to ship on time. We have pulled off miracles before, so that has now become the status quo. Our needs are rarely considered in the scheduling process, and it always kills us in the end. 3.5 General Conclusions The post-mortem demonstrated several critical issues: 1. Qualcomm is an engineering-driven company whose heart and soul is based on talented engineers allowed a certain level of creativity and freedom to innovate. Anything that would be perceived to break this culture would be aggressively opposed by management and employees alike.

23 2. The organization is fighting the transition to a more structured and mature state with standardized business practices. Most of the corporation understands the need for formal business processes and is actively looking to create them, while attempting to keep the underlying culture alive. 3. The biggest problems that were highlighted in the post mortems were caused by a breakdown of communications between the functional groups whose coordination is the job of program management. This disconnect between the groups has led to a mis-alignment of goals and expectations as to the launch of the product. 4. The control hierarchy design at the top of the pyramid and operations and marketing at the bottom has led to a sub-optimization of communication and management of programs. 4.0 The Balance Between Flexibility and Formality Throughout the NPI process there exists an inherent conflict between the culture of the design organization and the needs of manufacturing. Design represents creativity, independence and the lack of significant formality and structure. Dominant personalities within the organization have protected the autonomous environment of design engineering to the point where it is systemic throughout the company's culture and philosophy. Evidence of this can be seen by the lack of any design freeze throughout the NPI Process: At no point in the New Product Introduction process does the hardware and/or software design mature to the point where the engineering and manufacturing community recognize the product as stabilized and frozen. Though the revision control system at Rev - could be construed as a product freeze, in reality such measures are rarely implemented, and the design continues to evolve throughout the prototype, pre-production and production phases. However, this lack of formality is widely viewed as a organizational corecompetency, and something that should be protected at all costs. The contrasting needs of operations and manufacturing ultimately clash with this philosophy. Manufacturing is a pursuit that strives for repeatability, stability and precision. The ideal situation for manufacturing would be to have a finished and stable design with which to establish a supply chain and manufacturing process. The standard gating process to product development, where the design must be finished and frozen before moving onto the next step, is inherently manufacturing friendly. If the design is stable and well validated, finding suppliers and creating manufacturing processes to build the product is substantially easier. The trade-off is time-tomarket. However, waiting until the design is frozen to begin developing manufacturing

24 capability and a supply chain will lengthen the company's time-to-market, and make the product and company less competitive. In the highly competitive market of wireless telecommunications, one of Qualcomm's strongest competitive advantages has been time-to-market. As such, any recommendations that have been made to add structure to the NPI process have historically been quelled by engineering and management under the context that such formality would lengthen development cycle times. The quality organization is another internal group that historically pushed for more formality and structure to the products and processes. Quality assurance recognizes the risk of shipping product without more well established processes within the design phase, and has attempted to impose those beliefs upon the engineering community with gradual success. Up until the time of this thesis, Qualcomm had not encountered any major quality issues with products in the field, making it more difficult for the quality group to justify constraining checks and balances on the system. 4.1 NPI process - Critical Link The critical link in the NPI process is where the organizational culture of design engineering clashes with the organizational needs of manufacturing. This transition from phase 2 (Product Design) to phase 3 (Product Prototype) represents this critical point for Qualcomm's product introduction strategy. At this point, manufacturing is pulling into their organization as much information about the design and development as possible. That information is the base from which operations is establishing a short-term supply chain for the Prototype and Pre-production builds, long-term supply Time chains for the Production I,builds and establishment of their internal manufacturing Mock-up Prototype 0 Pre-Production capabilities and processes. Figure 6 Engineering's continuous iteration of the design through the Prototype and Pre-Production phases, forces operations to continually re-evaluate and change all work done up until that point. The sooner the majority of the iterations and changes in design are completed, the less potential impact they will have on manufacturing and the supply chain. Changes in the

25 ~bd r design that occur while manufacturing is in the prototype phase will impact the functional prototype and the short-term supply chain. Any changes in the design that occur while manufacturing is in the pre-production phase will impact not only the functional prototype and short-term supply chain, but also the pre-production units, the long-term supply chain and formalized manufacturing processes. Manufacturing would prefer to delay its actions as long as possible to allow the design to stabilize, but their ultimate deadline for product launch is rarely flexible. As such, manufacturing pushes up-stream into the design process in attempt to get information upon which to act, running the risk of having to adapt to massive changes in the design later on. The reality of technology development is that there will always be a number of changes that occur to the design well past the design phase, caused by customer concerns, hardware and software limitations and general market dynamics. Many of these influences cannot be controlled by the company at large. Though the engineering groups would like to "finish" the design before manufacturing get involved, time would not allow such a separation of tasks and inevitably changes will occur. These problems are not at all uncommon among organizations that practice concurrent engineering. 4.2 The Present Organizational Structure Qualcomm functionally organized with engineering and operations in two independent vertical structures. In the New Product Introduction process, the interaction and communication between the respective members of the engineering and operations Desig group is critical to the success of the NPI effort. As Materials described in the previous chapter, design engineering is perceived to be "in control," whereby other groups O ef M gineefing within the company must cater to design, pulling information from them in order to accomplish their own independent jobs. Recognizing the necessity to be Figure 7 close to the design engineers, the operations group has staffed the Engineering Planner and NPI Manufacturing Engineer in the same building as the engineering design team in hopes of closing the information gap between the functional groups. This has met with mixed success. The design

26 engineering groups tend to be very territorial, and the operations team members must work hard to foster long-term relations in hopes of gaining the trust and cooperation of the design engineers The Engineering Planner As presently defined in the organizational structure, the Engineering Planner is responsible for integrating changes in the product design, into the supply chain. This person's job begins as the design engineers start development on a product, and the design of certain electrical components or mechanical assemblies starts to take form. It is the responsibility of the Planner to drive the procurement and sourcing of the materials and components needed by the engineers. It is the responsibility of the Materials Planner to have visibility over all changes in the design as they occur, understand the implications to the supply chain and feed information back to the engineers on pricing and availability to the design engineer. Typically there are two Engineering Planners staffed on a given product, one specifically for the electrical Circuit Card Assemblies(CCAs) and one the mechanical assemblies and systems. Though these individuals are located in the same building as the design engineers, their offices are separated by approximately 200 yards and 2 flights of stairs The NPI Manufacturing Engineer The NPI Manufacturing Engineer (ME) is the liaison between the design engineers and the manufacturing community. The NPI ME is tasked with influencing the design for manufactureability. There is one NPI ME for each product being launched. This person is staffed in the same building the design engineering staff, and is primarily responsible for integrating operations with design engineering. They must keep abreast of any changes in the product design, understand their implication to the manufacturing process and capabilities, and communicate that information up-stream to the design team as well as down-stream to the manufacturing community. In phase 2 (Prototype) their job is to coordinate the development of the mock-up unit, and the subsequent Prototype units. As phase 2 comes to a close and Pre- Production is ramping-up the NPI ME is responsible for developing and documenting the preliminary manufacturing processes for production. During the Prototype and Pre-production phases of the NPI process, the Engineering Planner and the NPI ME are critical to the speed with which operations is able to setup the manufacturing capabilities for the product.

27 4.2.3 The Manufacturing Development Lab The Manufacturing Development Lab (MDL) is a 400 sq. ft. room off of the main manufacturing floor where the mock-up and prototype versions of the new BTS and BSCs are assembled. There are approximately 13 assemblers and technicians who work in the MDL and are dedicated to assembling the first functional units of any new product. Having the units created in the MDL allows manufacturing a certain amount of freedom, away from the constraints of shop floor controls, work instruction and specific routes. It also gives manufacturing the flexibility to rapidly change and manipulate the product as it takes form out of the design stages. At any given time, there can be up to four different products being built within this room. After there have been around 10 copies of the same product built in the MDL, the production is rolled out to the manufacturing floor with formally established and documented processes. 4.3 Definition of Optimization of the NPI critical Link Before making recommendations for the improvement of the NPI process or the critical link, a definition should be made as to what "improve" actually means in the context of New Product Introduction and specifically Qualcomm. Considering the contradictory cultures between design and manufacturing, the goal of an established NPI process is two fold. In one dimension, the NPI process should allow the engineers the creativity and independence that they desire, while giving manufacturing the formality and structure that they need. In the second dimension, the NPI process should be able to fully integrate the two functional groups and their respective attitudes to facilitate a quick transition from design to manufacturing. Considering Qualcomm's position in the wireless telecommunications market, the ultimate goal of the NPI process is to decrease the overall time to market of products for the Wireless Infrastructure Products Division (WIPD). 5.0 The Sync and Stabilize Model It is sometimes helpful to make comparisons between the organizational processes despite glaring differences in the products and markets. The Microsoft Corporation, based in Redmond Washington, and maker of software that runs on 95% of all personal computers, implements a model of design and development, called "Sync and Stabilize." 2 This method of organizing development teams is used extensively in software companies, and has been implemented effectively by Microsoft in the rapid development of large scale software programs. Though this